LSpineSMRI: A Comprehensive Dataset of Non-Contrast Lumbar Spine Stenosis MRI Examinations
Description
Central lumbar spinal stenosis (LSS) is a prevalent degenerative condition primarily affecting older adults, characterized by the narrowing of the lumbar spinal canal due to changes in intervertebral discs, ligamentum flavum, and facet joints. It is a leading cause of spine surgery in the elderly, leading to significant healthcare costs, especially when it results in severe pain and disability. LSS commonly presents as neurogenic claudication, with symptoms like lower back pain, leg discomfort, and tingling, which can severely impair mobility and quality of life. Diagnosing LSS requires a thorough evaluation of clinical symptoms and imaging findings, with magnetic resonance imaging (MRI) being the gold standard for visualizing soft tissues and nerve compression. This dataset includes imaging data and reports from consecutive non-contrast lumbar spine MRI examinations conducted between 2017 and 2023. It features a diverse patient population from inpatient, outpatient, and emergency services. Inclusion criteria required complete axial and sagittal T2-weighted sequences, all obtained using a 1.5 Tesla MRI scanner for consistent image quality. Exclusion criteria eliminated cases with severe spinal conditions such as trauma, infections, inflammation, tumors, and post-surgical complications that typically require gadolinium contrast. Patients with anatomical variations, fractures, extensive surgical hardware, or widespread metastatic disease were also excluded. MRIs lacking T2-weighted axial sequences or with poor quality were omitted to ensure dataset reliability. Ultimately, 640 patients (234 males and 406 females) met the criteria, with a mean age of 57.58 ± 12.47 years (range: 21 to 87). A total of 17,440 MRI slices were collected and categorized into cervical/thoracic, lumbar, and sacral regions, with lumbar slices subsequently graded. This dataset received approval from the Iran University of Medical Sciences Ethics Committee. Qualitative grading was performed by a board-certified neurosurgeon with over 20 years of experience. The grading system is as follows: Grade A indicates no or minimal stenosis with visible cerebrospinal fluid (CSF); Grade B denotes moderate stenosis, where rootlets occupy the dural sac but remain distinguishable; Grade C signifies severe stenosis, with a homogeneous gray signal in the dural sac and no recognizable rootlets or visible CSF, although posterior epidural fat is present; and Grade D reflects extreme stenosis, with neither rootlets nor posterior epidural fat visible. In this version, Grades B and C have been merged. Annotations were performed using LabelStudio, a versatile tool for various data types, including images and text. This dataset serves as a valuable resource for understanding lumbar spinal stenosis and has potential applications in developing deep learning-based computer-aided diagnosis (CAD) systems, enhancing medical imaging and diagnostic accuracy.